The Suda On Line: Applying Computer Technology to Ancient and Byzantine Studies

“The Suda On Line: Applying Computer Technology to Ancient and Byzantine Studies”

Ross Scaife U of Kentucky scaife@uky.edu Raphael Finkel U of Kentucky raphael@cs.uky.edu

Using highly interdisciplinary methods we have built a collaborative infrastructure for translation and annotation of ancient texts. This generalizable infrastructure is now fully deployed in the Suda On Line http://www.stoa.org/sol/). The Suda is a 10th century Byzantine Greek lexicon of some 30,000 lemmata. After four years of continuous development we have implemented a complex yet effective and practical system. Our goal is not only to provide the SOL as a useful tool for researchers, but also to explore and facilitate the modes of scholarship now made possible by open source technology and the internet: this effort is cooperative rather than solitary, communal rather than proprietary, worldwide rather than localized, evolving rather than static. Our international team of managing editors, editors, and translators has now worked up approximately one third of the material in the Suda, quite a satisfactory rate of progress. ACH/ALLC in Glasgow had an initial presentation concerning this project; we feel that substantial further development and our positive results warrant an update at this time. In order to encourage the participation of translators and editors, and in order to make the SOL database a useful scholarly resource as quickly as possible, we make our materials available to users as soon as it is submitted. We acknowledge that this philosophy raises concerns. One of the major issues with electronic publication of scholarship is the potential it has for circumventing time-tested procedures for quality control. While we do not want simply to add to the sea of uncontrolled material on the Web, at the same time we insist on our right to experiment, and we have no desire to replicate the print-publication paradigm in electronic format. Many of the advantages that electronic publication offers, including immediacy, accessibility and adaptability, are seriously handicapped by traditional editorial processes, where chronic bottlenecks frequently develop in the effort to keep the publishing house’s imprimatur off of anything with any detectable shortcomings. In order to exploit these advantages of the web while at the same time maintaining a reasonable level of quality control, submissions to the SOL database undergo the following process of editorial evaluation and monitoring:

1. Initial submissions immediately become accessible to users searching or browsing at the SOL site, but their “draft” status is clearly marked.
2. Once a submission has been carefully vetted by one of the SOL editors for errors and significant omissions, its status as part of the SOL database may rise from draft into one of two categories: low or high. At every stage of this process, the editors who participate in vetting and improving the entry will be prominently identified to the user, along with any descriptive comments they may provide concerning their editorial work.
3. Most importantly, even an entry that has achieved high status will not be considered perfect and immutable. At the discretion of the editors, improvements, changes and additions of links and bibliography can continue indefinitely.

While this way of doing things puts more of the burden of quality control on the end user, our system of marking editorial status gives researchers significant assistance in coming to an informed decision about the reliability of the material in SOL. In fact, our system offers definite advantages over the canonical paradigm of peer review from the consumer’s point of view. In print scholarship (and electronic scholarship that merely follows the traditional model) the number, identity, and qualifications of reviewers remain hidden, and one must usually base one’s estimate of the reliability of the scholarship solely on the identity of the author and the general reputation of the venue. In the standard paradigm, moreover, the end product is more or less fixed, whereas our database is being improved continuously. This presentation will describe our project from various perspectives, including the following principal points of discussion. (1) An overview of the Suda itself, including a few examples that illustrate its diverse composition and unique value for several fields of humanistic scholarship, despite its flaws and peculiarities. (2) The multiyear interdisciplinary collaboration among computer scientists, historians, and philologists that has produced our results so far. (3) The academic ideology that guides our production of a freely-available and open-ended e-text, including significant ways in which our editorial practices diverge from more traditional ones. (4) The most important features of the online site available on a hierarchical basis to the participants and the general public. (5) The specific applications and programming technologies that enable those features. (6) Our most recent effort: generation of a unified, complete, and self-documenting XML snapshot of our data. This latest ability addresses our responsibility to ensure the long-term archival security and viability of our results, and it also allows us to experiment with powerful new technologies centered around XSLT programming and the Cocoon environment for the transformation and publication of electronic documents. The presentation will include a demonstration of these experiments and conclude with the prospects for future developments.